There are needs to support hidden/restricted data. Investigate what we might do

[deanwampler]

Two classes of data that shouldn't be open sourced:

1. Benchmark data that people don't want to be "vacuumed" into training data sets.
2. Data with export or other security restrictions.
3. Private organizational data.

While our goal is to support open datasets, should we try to address these needs?

[blublinsky]

Currently Hugging Face already supports the notion of gated datasets. Here is their definition:

To give more control over how datasets are used, the Hub allows datasets authors to enable access requests for their datasets. Users must agree to share their contact information (username and email address) with the datasets authors to access the datasets files when enabled. Datasets authors can configure this request with additional fields. A dataset with access requests enabled is called a gated dataset. Access requests are always granted to individual users rather than to entire organizations. A common use case of gated datasets is to provide access to early research datasets before the wider release.

Will this satisfy the requirement?